Cadillac Opens Its F1 Engine Factory on 18 December, but the Real Question Sits Between 2029 and 2031

Across the last four qualifying sessions of the 2026 season, Cadillac has been the slowest car on the grid. The team sits on zero points — the only one of eleven without a score. And on 18 December, that same team opens a power unit factory in Charlotte.

Taken together, those two facts are a data anomaly. The team that has been last on the grid for four consecutive rounds is putting capital into building the cause of a lap time it currently cannot produce. For me the centre of this story is not the date. The centre is that 18 December is a civil-works milestone, not a technical one — the announcement carries not a single figure for dyno cells, combustion test rigs, ERS benches or engineering headcount.

Cadillac entered Formula 1 in 2026 as the eleventh team, chassis designation MAC-26. Until 2028 the team runs customer Ferrari power units; the FIA approved GM as a power unit supplier from 2029, with that approval granted in 2026. GM president Mark Reuss confirmed the factory date at the Autosport Business Exchange and added that the build is 'logarithmic, not linear' — an early effort whose output will not show up immediately. He held up Cadillac's GTP/hypercar programme as the precedent, a case of building a competitive package from zero. The driver line-up is deliberately experienced — Sergio Perez and Valtteri Bottas — and both are currently recording the worst results of their careers.

Then there is the regulatory weather. FIA president Mohammed Ben Sulayem has publicly committed to a V8 return by 2031, with a signal toward sustainable fuel carrying a smaller electric share. That is not a calendar tweak; it is a partial reversal of the technical route the 2026 rules were built around.

Start with the part no number covers. A factory date tells you nothing about a power unit's strength. Real capability is measured in three places: the count of dyno cells, the availability of combustion and ERS test rigs, and the number of engineers who can stand at them. Not one of those figures is public. So I will file 18 December as a facilities milestone, not technical progress.

Second, the customer power unit ceiling. Until 2028 Cadillac runs Ferrari's unit, which means no control over packaging, cooling architecture or energy deployment mapping. This is an established paddock pattern and the price is paid twice — in lap time and in setup freedom. A team that cannot define its own cooling window has already lost half a Friday.

Third, the chassis verdict. Being slowest in four consecutive qualifying sessions does not mean the team is fighting someone and losing. It means the team is not fighting anyone — it is racing the clock. The wind tunnel and CFD numbers are not correlating with the track, and that gap is structural, not marginal. In thirty-one years of race weekends the pattern with new teams is consistent: the first season builds a measurement system, not a car.

Here I put a model on the table, because a model has no name until you build it yourself. Call it the Works Window Index. The definition is plain — how many seasons a manufacturer gets from its power unit, divided by a normal amortisation window. Take six seasons as normal. GM's unit homologates for 2029; if V8 arrives in 2031 that is two seasons, and if it slips to 2030, one. The index reads 0.33 or 0.17. Stated plainly: this is a ratio, not a probability. I have three historical cycles to compare against (2026 hybrid, 2026-2026, 2026-2026), so n=3 — from a sample that small I will not claim a likelihood, only a scale.

Fourth, the character of the 2029 unit. Under a compressed timeline and 2031 uncertainty, a first-generation engine usually prefers reliability to performance. My expectation is that GM's first unit will be a finisher rather than a benchmark.

And the line that must be printed every time: this model cannot see who is standing at the rig at two in the morning in Charlotte, whose gardening leave ends when, or how many years of patience the board actually holds. The spreadsheet held more truth than the highlight reel that week — but a spreadsheet cannot measure morale.

After 18 December the headlines will assume a causal link that does not exist. Even correlation between the factory opening and lap time improvement is currently zero; causation is far away. Reuss's 'logarithmic, not linear' line is an accurate engineering description — and also expectation management, because it pre-emptively delinks the factory opening from any near-term result.

My habit is to audit the paddock consensus too. The claim that the 2026 regulation reset was the correct entry window is one I agree with; that is a tally on my own scorecard. But the assumption that the cost cap helps a new team is aimed at the wrong target. The cap compresses what big teams can spend; it does not build a new team's infrastructure. Cadillac's binding constraint is time and people, not money.

The third contrarian angle: Cadillac's exposure is regulatory direction, not compliance. The FIA can reshape the technical formula on a timescale shorter than a manufacturer's development cycle. That is not new in this sport, but for every entrant it is a hidden tax on capital — and for the other ten teams too, since an eleventh team's share comes out of the same prize-money pool.

Three numbers to watch from here: whether the V8 ratification lands in 2030 or 2031; whether GM announces a customer team for its 2029 unit (none yet, which is itself the signal that the programme is not commercially mature); and when Charlotte publishes a dyno cell count. A driver line-up review around 2028 is close to certain — Reuss's 'right people, right engineers and right drivers' is not a sentence about anything else.

The question is this: if a manufacturer gets two seasons out of a power unit and the rules change three years later, does it knock on the door a second time?